Research on Academic Early Warning Model Based on Improved SVM Algorithm

Abstract

In this paper, machine learning technology is applied to the study of student academic early warning, and a student academic early warning model is constructed to help college education managers fully understand students, accurately predict students and personalized service students. Aiming at the problem that the SVM algorithm uses the default value of the penalty factor γ and the kernel parameter gamma when constructing the academic early warning model, the prediction model cannot achieve higher accuracy, a SVM algorithm based on improved FOA is proposed. Based on the first three years ' score data and library data of a university student, the SVM algorithm based on improved FOA is used to predict whether students can graduate smoothly in the future, and to give academic warning to students who may not graduate smoothly in the future. Experiments show that the SVM early warning model based on improved FOA is superior to the three types of traditional SVM model, decision tree and random forest in terms of accuracy.